An AI-first knowledge system that lets users capture notes and ask their own knowledge through grounded, source-backed AI responses.
This project focuses on clean architecture, security, and disciplined AI usage, inspired by Mem.ai.
Ask your knowledge, don’t manage it.
Users write freely. The system organizes and retrieves information using AI—without hallucinations.
- JWT + Google OAuth authentication
- Notes and collections CRUD (user-scoped)
- Async AI summarization and auto-tagging
- Grounded chat with notes (answers only from stored data)
- Source-attributed AI responses
- Token-efficient AI usage
- Simple, calm, responsive UI with subtle motion
- Frontend: Next.js (App Router), React, Tailwind CSS, Framer Motion
- Backend: Next.js API Routes (Node.js)
- Database: PostgreSQL (Supabase)
- Auth: JWT (stateless) + Google OAuth
- AI: Gemini API (server-side only)
- Uses only user-provided notes
- Returns “I don’t have enough information” when context is insufficient
- Summaries and tags are cached
- AI never blocks user actions
- JWT validation on all protected routes
- Strict user data isolation (
user_idscoping) - Passwords stored separately (
user_secrets) - No client-side AI calls
- No sensitive data exposed
Core functionality complete.
Advanced features (vector search, public APIs, collaboration) intentionally excluded.
Manoj Mannam